sentiment_trading

Inspiration

We like competitions

What it does

Tries to beat other market makers by

• Running sentiment analysis model on news
• Taking an aggressive order with a new target price deduced by the sentiment and then stay out of the market for a few seconds (don't cancel orders)
• Check cross exchange spread for arbitrage opportunities
• Guardrails for soft position limit -> rebalancing if needed

file: sentiment_analysis/run_trading_bot.ipynb

How we built it

Optibook ENV + transformers library for sentiment analysis + some pure python code for determining strategies

Challenges we ran into

Understanding desired price movements, trying to improve sentiment analysis model

Accomplishments that we're proud of

Finally making a profit

What we learned

Intuition is not enough, trading is hard and knowledge is key

What's next for sentiment_trading

Built With

• python
• transformers